Dohyoung Kim

Mr. Dohyoung Kim

orcid

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
26 November 2024 in Electronics
Verified
Journal Article
Open access
PDF file
DKDohyoung Kim
Dohyoung Kim
HWHyekyung Woo
Hyekyung Woo
YLYoungho Lee
Youngho Lee
The rapid increase in data volume and variety within the field of machine learning necessitates ethical data utilization and adherence to strict privacy protection standards. Fair federated learning (FFL) has emerged as a pivotal solution for ensuring fairness and privacy protection within distributed learning environments. FFL not only enhances privacy safeguards but also addresses inherent limitations of existing federated learning (FL) systems by fostering equitable model training across diverse participant groups, mitigating the exclusion of individual users or minorities, and improving overall model fairness. This study examines the causes of bias and fairness within existing FL systems and categorizes solutions according to data partitioning strategies, privacy mechanisms, applicable machine learning models, communication architectures, and technologies designed to manage heterogeneity. To mitigate bias, enhance fairness, and strengthen privacy protections in FL, this study also explores fairness evaluation metrics, relevant applications, and associated challenges of FFL. Addressing bias, fairness, and privacy concerns across all mechanisms serves as a valuable resource for practitioners aiming to develop efficient FL solutions.

19 June 2025 in Expert Systems with Applications
Verified
Journal Article
Open access
DKDohyoung Kim
Dohyoung Kim
경오경수 오
경수 오
YLYoungho Lee
Youngho Lee
HWHyekyung Woo
Hyekyung Woo

29 January 2025 in IEEE Access
Verified
Journal Article
Open access
DKDohyoung Kim
Dohyoung Kim
KLKangYoon Lee
KangYoon Lee
YLYoungho Lee
Youngho Lee
HWHyekyung Woo
Hyekyung Woo
One of the major drawbacks of federated learning (FL) is data imbalance and uneven reliability of clients, which adversely impacts model performance and generalization ability. Among these problems, data imbalance can be addressed through semi-supervised federated learning (SSFL), in which fully labeled, and fully unlabeled clients are jointly trained to derive a global model. Existing approaches work well when local clients have independent identically distributed (IID) data but fail to generalize under a more practical SSFL setting, namely fairness, non-IID. Since the non-IID problem stems from unfairness, to address this issue, we propose ACMFed, a fair and improved concurrent federated learning method that balances client connections, computing resources, data quality and quantity based on fairness. ACMFed reaches the compromise by performing fair multiple sampling of clients in models with largely deviating labeled and unlabeled clients. Each client distills the model through a multi-sub-sampling process without aggregating the local model to the central server, and builds a global model based on it. The additional compromise model applies a fair adjustment method of the new distance-based dynamic model to reduce data bias and optimize performance. ACMFed outperforms other federated learning methods and achieves faster convergence on six benchmark datasets. To code is available at https://github.com/kimdohyoung96/ACMFed.

29 February 2024
Verified
Book Chapter
Closed access
PDF file
DKDohyoung Kim
Dohyoung Kim
경오경수 오
경수 오
SKSeok-Hwan Kang
Seok-Hwan Kang
YLYoungho Lee
Youngho Lee

15 October 2024 in Preprints.org
Verified
Preprint
Open access
PDF file
DKDohyoung Kim
Dohyoung Kim
HWHyekyung Woo
Hyekyung Woo
YLYoungho Lee
Youngho Lee
In the field of machine learning, the rapid development of data volume and variety requires ethical data utilization and strict privacy protection standards. Fair Federated Learning (FFL) has emerged as a key solution that aims to ensure fairness and privacy protection in a distributed learning environment. FFL enhances privacy protection and solves the inherent limitations of existing federated learning (FL) by promoting fair model training in diverse participant groups, preventing the exclusion of individual users or minorities, and improving overall model fairness. In this study, FFL discusses the causes of bias and fairness of existing FL, and separates solutions based on data partitioning strategies, privacy mechanisms, applicable machine learning models, communication architectures, and technologies to overcome heterogeneity. In order to improve the causes of bias, fairness, and privacy protection of FL, fairness evaluation indicators and applications and challenges of FFL are discussed. Since it addresses bias, fairness, and privacy issues in FL of all mechanisms, it can be an important resource for practitioners who want to implement efficient FL solutions.

28 February 2025 in Journal of Health Informatics and Statistics
Verified
Journal Article
Open access
PDF file
KNKyungmin Na
Kyungmin Na
DKDohyoung Kim
Dohyoung Kim
YLYoungho Lee
Youngho Lee
Objectives: This study aims to compare the performance of federated learning (FL) and fair federated learning (FFL) in classifying pneumonia patients based on chest X-ray data. The primary focus is on assessing the accuracy and fairness of these models in handling imbalanced and distributed data in real-world healthcare settings.Methods: We used a large chest X-ray dataset to evaluate the performance of FL and FFL models. The models were built using the ResNet50 architecture, and experiments were conducted under both independent and identically distributed (IID) and non-IID data conditions. The FFL approach applied optimized loss functions to address data imbalance and ensure fair contribution from each client, regardless of the local data distribution.Results: Our findings indicate that FFL consistently outperforms traditional FL models, particularly in non-IID environments. The FFL model demonstrated higher accuracy in pneumonia classification, achieving a significant improvement in model fairness and performance across different client datasets. The use of the ResNet50 architecture further enhanced the model’s ability to handle complex X-ray image patterns.Conclusions: FFL offers a superior solution for handling imbalanced medical data compared to conventional FL models. Its ability to maintain fairness while improving classification accuracy makes it an ideal approach for decentralized healthcare systems, ensuring better patient outcomes while preserving data privacy.